169 alcohol (ca. 10 mi) was added and evaporation and cooling caused precipitation of the Reissert compound. Treatment and of the Reissert compounds ( x g) dissolved in a minimum of hot acetic acid, with 40% aqueous fluoroboric acid (3.5 X x g), followed by cooling produced the hydrofluoroborate salt which was collected by filtration and either washed or recrystallized from ethanol. Reissert compounds and their salts obtained via this general procedure are reported in Table 111. New compounds gave satisfactory elemental analysis data and melting points of known compounds agreed with those reported previously. a-Deuterated Reissert Compounds. These were prepared by the following general route. The Reissert compound (2 g) was taken up i n sufficient TFA to obtain a clear solution (ca. 4 ml). This solution was added dropwise to a swirled flask containing dry ether (100 ml). The solid was filtered off and ground with anhydrous sodium carbonate in a mortar and pestle. DzO (5-10 ml) was added to obtain a thick paste which was mixed thoroughly for 5 min. The mixture was then taken up in chloroform (20 ml), the organic layer separated, and the aqueous layer washed with chloroform (2 X 10 ml). The combined organic fractions were then dried (NaZC03) and evaporated to dryness, and the product was recrystallized from 1 : I O benzene/cyclohexane (ca. 20-40 ml). The a-deuterated Reissert compounds were shown to have melting points and mixed melting points identical with the parent compounds.
Acknowledgments. We thank Professor W. E. McEwen for helpful discussion and the S.R.C. for financial support (to A.D.P.).
References and Notes (1) (a) Part XII: S. 0.Chua, M. J. Cook, and A. R. Katritzky. J. Chem. SOC.6, 2350 (1971); (b) part XIII: G. Bianchi, A. J. Boulton, I . J. Fletcher, and A. R. Katritzky, ibid., 2355 (1971). (2) For reviews, see (a) W. E. McEwen and R. L. Cobb, Chem. Rev., 55, 5 1 1 (1955); (b) F. D. Popp, Adv. Heterocycl. Chem., 9, 1 (1968). (3) See, e.g., F. D. Popp and W. Blount, J. Org. Cbem., 27, 297 (1962). (4) R. L. Cobb and W. E. McEwen, J. Am. Chem. SOC.,77,5042 (1955). (5) J. W. Davis, Jr., J. Org. Chem., 25, 376 (1960). (6) W. E.McEwen. I. C. Mineo, and Y. H.Shen, J. Am. Chem. Soc., 93,4479 (1971). (7) For reviews, see (a) A. R. Katritzky and J. M. Lagowski, Adv. Heterocycl. Chem., 2,27 (1963); (b)J. Elguero, C. Marzin. A. R. Katritzky, and P. Linda, “The Tautomerism of Heterocycles”, Academic Press, New York, N.Y., 1976. (8) See ref 7b, pp 431-433. (9) W. E. McEwen, M. A. Calabro, I. C. Mineo, and I . C. Wang. J. Am. Chem. SOC.. 95, 2392 (1973). (10) A. J. Birch, A. H. Jackson, and P. V. R. Shannon, Tetrahedron Lett., 4789 (1972). (11) J. M. Bobbitt. J. M. Kiely. K. L. Khanna, and R. Ebermann, J. Org. Chem., 30, 2247 (1965); J. M. Bobbitt and J. C. Sih, ibid, 33, 856 (1968). (12) A. D. Page, Ph.D. Thesis, University of East Anglia, 1976. (13) F. 0.Popp, K. T. Potts, and R. Armbruster, Org. Mass Specrrom., 3, 1075 (1970). (14) A. Bischler and B. Napieralski, Chem. Ber., 26, 1903 (1893). (15) W. M. Whaley and M. Meadow, J. Chem. Soc.,1067 (1953). (16) Moreover. no authenticated examples of the isolation of a nonthermodynamically stable imino form are known.’ (17) Examinatlon of the actual spectra as recorded in the Ph.D. Thesis of I. C. Wang, University of Massachusetts, 1971, supports this view.
Kinetics and Thermodynamics of Intermolecular Saturated Amine Excimers Arthur M. Halpern,*’ Pierre Ravinet,2 and Ronald J. Sternfels3 Contribution from the Photochemistry and Spectroscopy Laboratory, Department of Chemistry, Northeastern University, Boston, Massachusetts 021 15. Received April 30, 1976
Abstract: The rate constants and activation parameters pertaining to excimer formation in 1-azabicyclo[2.2.2]octane(ABCO) and 1-azaadamantane (1-AA) have been measured in n-hexane solution utilizing fluorescence decay data. These excimers are 375 and 395 nm, respectively). The formation characterized spectroscopically by broad, very red-shifted emission bands (A,, rate constants are diffusion controlled and have been analyzed without the need to consider the “transient term”. These excimers are rather tightly bound having AH values of - I 1.6 and -10.2 kcal/mol, respectively. In addition, the dissociation rate constant preexponential factors are typically large, with AS values of ca. 19 and 17 eu, respectively. The methodology of determining binding energies using the steady-state method vis-a-vis kinetic procedures is discussed; the former technique provides lower limits of 1 AH1 values. For ABCO, the excimer relaxation rate constant, k D , was found to be slightly temperature dependent ( W D 1.4 kcal/mol), and an analysis of the temperature dependence of the excimer fluorescence quantum intensity revealed that it is the nonradiative portion of kD which varies with temperature. The ABCO excimer is weakly quenched by ground-state amine ( k Q D = 7.2 X lo7 M-’ s-’). The photophysical properties of nonexcimer-forming amines are compared with ABCO and I-AA, and an attempt is made to rationalize the stability of these excimers within simple exciton and charge resonance models. Neither approach appears to be adequate in accounting for the observed binding energies or in explaining the structural specificity in excimer formation in saturated amines. The structure of the excimer is discussed and configurations are proposed: head-to-head (N-N) and C-N bond antiparallel to C-N bond. The head-to-tail arrangement is excluded because 4-Me-ABCO also undergoes excimer formation and possesses similar thermodynamic properties. The photochemistry of ABCO and triethylamine (in n-hexane) solution is also discussed. N
Introduction The association between an electronically excited molecule and its ground-state counterpart (Le., excimer formation) is a well-known phenomenon which has been extensively studied, both experimentally and theoretically, for about 20 year^.^-^ Until rather recently, however, investigations of organic molecular excimers and their formation, deactivation, stability, etc. have been confined to aromatic or heteroaromatic moleHalpern et al.
c u l e ~ .Studies ~ - ~ of excimer formation in some nonaromatic compounds, Le., saturated tertiary amines, have been described r e ~ e n t l y . It~ ~thus ’ ~ appears that the presence of the p-electronic system characteristic of aromatic molecules is not an essential criterion for the stabilization of excimers in organic molecules.” Thus the nonbonding electron pair in a saturated amine, when properly oriented in certain aliphatic systems can be very important in effecting excimer stabilization both intermolecularly and intramolecularly.’ It is the purpose of this
/ Kinetics and Thermodynamics of Intermolecular Saturated Amine Excimers
170 paper to present the results of detailed kinetic and thermodynamic studies of some saturated amine excimers, to suggest possible geometries of these interesting systems, and to speculate about the nature of the binding forces in these excimers. Although the photophysical properties of a large number of saturated amines have been investigated in this laboratory under widely varying experimental conditions (Le., vapor phase, solvent media of differing polarity, temperature, etc.), only three amines have been unequivocally demonstrated to undergo intermolecular excimer formation. These compounds are: l-azabicyclo[2.2.2]octane (ABCO), 4-Me-ABCO, and
F‘
200
I
,’!,
I
I
I
280
A,
I
I
I
440
360 NM
Figure 1. Spectra of ABCO in n-hexane solution at 23 “C: (-) absorption (-*-) 2.0 X M; spectrum; (- - - -) emission spectrum 2.0 X 2.0 X IO-* M. Emission spectra are uncorrected and are obtained with A,, of 238 nm. ( . . . a
1A A
ABCO 1-azaadamantane (1 -AA). The vapor pressures of ABCO and 4-Me-ABCO are sufficiently high under ambient conditions so that excimer formation is observable in the vapor phase for these amines; for example, the vapor pressure of ABCO is about 2 Torr (ca. 1 X M) a t 23 O C and is large enough to allow pressure dependent studies to be carried out. The formation and dissociation kinetics of the ABCO vapor phase excimer have been reported recently.I0 The vapor pressure of 4-Me-ABCO is about 3 Torr under ambient conditions.
Results and Discussion Absorption Spectra. Unless otherwise indicated, all studies were carried out using n-hexane as the solvent medium. First the electronic absorption spectra of the amines will be summarized. Assignments of the electronic absorption spectra of the trialkylamines have been suggested only in the vapor phase wherein the upper states are believe to be Rydberg in nature.I3 Thus, the lowest lying excited state is assigned as (n,3s) and the next higher two states as (n,3p) and (n,4p). Although it is recognized that states arising from valence shell conjugate orbitals (i.e., cr* orbitals) are important, the location of these V N states in the trialkylamines has not been conclusively identified.I3.l4 The vapor phase absorption spectra of each of the excimer-forming cage amines depicted above are characterized by highly structured vibronic features.’ 5,16 Significantly, these vibrational patterns are completely absent in the spectra obtained in an n-hexane medium, This phenomenon can be interpreted as being a consequence of the Rydberg nature of the excited states.13 It is probably incorrect, however, to assume that in condensed media, the Rydberg states become primarily “washed out” and replaced by the valence shell states. In any event, assuming for the moment that the “integrity” (and/or importance) of the Rydberg states is diminished in the condensed phase relative to the vapor state, one is led to the conclusion that the electronic character (i.e., the degree of Rydberg/valence shell importance) of the upper state(s) is not crucial to excimer formation in the amines. This follows from the fact that the energetics and kinetics of excimer formation in ABCO are not very different in the vapor phase vis-a-vis n-hexane solution.1° The absorption spectra of the trialkylamines in n-hexane solution (including the cage amines discussed here) are rather undistinguished, consisting of a moderately intense maximum in the 200-nm region with tailing to lower energies, out to ca. 255 nm. In certain amines, e.g., ABCO, an inflexion in the 230-nm region can be discerned, and using the more infor-
-
Journal of the American Chemical Society / 99:l
e)
mation-laden vapor phase spectra as a guide, the transition possessing the maximum near 200 nm can most likely be correlated with the 3p n transition (SZ SO),and the long wavelength shoulder (or tail) can be associated accordingly with the 3s n transition (SI-SO). It should be realized that the presence of the valence shell conjugate as a continuum overlapping these (proposed) Rydberg excitations cannot be ignored. The detailed vapor phase absorption spectra of ABCO have been reported e l s e ~ h e r e , and ’ ~ the 3p n (SZ SO) spectrum of 1-AA has been published.16 The absorption spectrum of ABCO in n-hexane is shown in Figure 1, and as mentioned above, the rich vibronic structure which characterizes the vapor state spectra is completely absent. Thus this spectrum resembles those of other saturated amines.” It is important to note that the cage amines reported in this paper, as well as many other monocyclic and acyclic amines, obey Beer’s Law in hydrocarbon solvents up to a t least ca. IO-’ M). This seems to preclude the possibility of (discrete) ground-state interactions between these saturated amine solutes. Fluorescence Spectra. In the vapor phase, both ABCO and 4-Me-ABCO are characterized by rather highly structured fluorescence spectra (if the excitation energies are kept below ca. 4000 cm-’ of the respective S I SOenergies (and/or vibrational relaxation is ensured by the presence of a buffer gas). ABCO’s fluorescence spectrum has been described elsewhere. 10,18In n-hexane solution, the fluorescence spectra, like the absorption spectra of the three cage amines reported here are structureless, having maxima in the 270-286 nm region. Some pertinent data concerning the absorption and fluorescence spectra of ABCO, 4-Me-ABCO, and 1-AA are summarized in Table I. These data show that in refluorescence, there is no discernible shift in the solution phase maxima when compared with the (estimated) maxima of the respective Franck-Condon envelopes of the vapor phase spectra. It should also be noted that the transition energies of ABCO and 4Me-ABCO are very similar whereas 1-AA absorbs (and emits) considerably to the red, presumably as a consequence of the higher degree of alkylation (and hyperconjugation) of the latter amine.19 The lack of a dramatic solvation effect on the fluorescence spectra of these amines is somewhat puzzling in view of the presumed Rydberg nature of their upper states. On the other hand, it has been suggested that the fluorescence from simple acyclic amines originates from an underlying valence shell state.20This point is not easily disposed of, because in the cage amines, the correspondence between the fluorescence and S I SOabsorption spectra in the vapor phase (Le., the coincidence of the 0-0 bands, correlation of vibrational progression frequencies, etc.) would seem to support the idea that fluorescence originates in fact from the ( n ~ , 3 s Rydberg ) state. It is interesting to compare the energies of the fluorescence
-
-
+-
- -
-
-
/ January 5, 1977
171 Table I. Absorption and Fluorescence Data for the Cage Amines in the Vapor Phase and in n-Hexane Solution at 23 "C Absorption
SI
Amine ABCO 4-MA 1-AA
-
FluorescenceU
Vapor
SO
Sz
-
SO
228.0 227.8 236.7
255.9 252.3 265.5
Solution
S3 -SO
, , A,
(emaJC
206 (1980) 207 (2330) 211 (1970)
188.3
e 195.8
Vapor
e (238 nm)c
Monomer
Excimer
Monomer
Excimer
247 222 252
27Id 268 289d
375 375
272.5 270 286.5
375 375 395
Wavelength values in nm. 0-0 bands. M-' cm-l. Emission is structured; apparent, ,A, is not located with certainty. f Excimer emission is not observed in the vapor phase at 23 O C .
5 c
34
36
30
40
42
maxima of ABCO (or 4-Me-ABCO) and an acyclic amine containing a similar number of alkyl groups, e.g., triethylamine. A,, (fluorescence) for triethylamine lies about 10 nm to the red of ABCO, Le., a t about 282 nm (see Table I). This is probably due to the fact that whereas emission from the acyclic amine takes plce from a planar (or nearly planar) equilibrated excited state, the three ethylene bridges constrain ABCO from achieving a planar upper state, and this molecule will exhibit, accordingly, less of a geometry change-inducing Stokes shift that flexible triethylamine. These structural and spectroscopic distinctions between rigid and nonrigid amines bear interesting consequences vis-a-vis the temperature dependence of fluorescence intensity in these types of amines.*' The fluorescence spectra of all nonrigid (tertiary) amines examined to date are concentration independent, although many of these amines undergo self-quenching. Many rigid amines, such as I-azabicyclo[3.3.3]undecaneand 1,4-diazabicyclo[2.2.2]octane (DABCO), also exhibit concentration independent fluorescence spectra. On the other hand, the spectra of ABCO, 4-Me-ABCO, and 1-AA are all concentration dependent. Thus in dilute solutions (ca. 10-5-10-4M in n-hexane), these amines fluoresce in the ultraviolet region and, as such, resemble typical saturated amines (except for a small blue shift, vide supra). As the concentration of these amines is increased, a significantly red-shifted, broad, structureless emission band appears and gains prominence with increasing concentration. At about M, the fluorescence spectra of these cage amines are almost fully characterized by the respective long wavelength bands, and the fluorescence is perceived by the eye as a bright violet color. This behavior is also illustrated in Figure 1 for ABCO. The fluorescence Halpern et al.
Solution
f
of Franck-Condon envelope. The origin
maxima of the long wavelength bands of the cage amines are summarized in Table I. This long wavelength emission band has been assigned as excimer fluorescence on the basis of its temporal, thermal, and concentration dependence. The short wavelength fluorescence band (the only band observable in very dilute solutions) will be referred to as monomer emission. It should be noted that the series of emission spectra shown in Figure 1 for ABCO does not exhibit the usual isoemissive point which characterizes most aromatic excimer systems. The failure of ABCO to manifest an isoemissive point is a consequence of excimer quenching by ground-state molecules and will be discussed in more detail below. First, the temperature dependence of the monomer and excimer bands will be discussed, and the results of kinetic studies will be presented subsequently. It will be convenient,
huM
+A
?
?
2A
+ hull
tems. The denotation of rate constants follows that of Birkss Monomer and excimer decay constants are defined as follows: kM kD
= kiM
= kiD
+ kfM
kfD
i- ~ o D [ C I
where k i M and k f M , k i D and k f D are the nonradiative and radiative rate constants for the monomer and excimer, respectively. One additional step has been added to the usual monomer/excimer system, namely, the quenching of the excimer by ground-state amine (rate constant k Q D ) . Under conditions of steady-state excitation, the well-known relation for the ratio of excimer to monomer fluorescence intensities is:
where DM and k M D are the excimer formation and dissociation rate constants, respectively; k M and k D are the monomer and excimer composite relaxation rate constants irrespective of the coupling with each other; k D is concentration dependent as a consequence of the quenching process mentioned above. In eq 1, k D M and k M D are usually regarded as being the only
/ Kinetics and Thermodynamics of Intermolecular Saturated Amine Excimers
172
1,
O C
‘
O
T
where the A’s and Ws are the respective activation parameters for excimer formation and di~sociation.~ Although knowledge of the value of T,,, obtained experimentally is not very useful in determining any of the rate or activation parameters per se, eq l a can be used to check the internal consistency of the
various activation parameters once determined by other, more direct means. Experimentally measured T,,, values for ABCO and 1-AA are 8 and -10 OC, respectively, and are favorably compared with the calculated values of 1 4 and -13 OC for these systems. It should be noted that the slight temperature dependence observed for kD in the ABCO excimer vide infra renders eq 1a approximate; nevertheless, the decent agreement between the measured and calculated T,,, values implies that the approximation is reasonably good in these cases.
Figure 3. Plot of In [ I r o / l r M ] vs. 1 / T for I -AA in n-hexane (in the hightemperature region); A,, is 238 nm.
(strongly) temperature-dependent rate constants.Z2 If the temperature is high enough so that the excimer dissociation rate exceeds the intrinsic (Le., radiative and nonradiative) depletion rate of the excimer, k M D > k D , and an Arrhenius plot of the ratio of excimer to monomer fluorescence intensities can be used to obtain approximate values of the enthalpy and entropy of excimer formation.23In the “low-temperature region”, ~ M < Q k D , and an Arrhenius plot of eq 1 can provide an estimate of the activation energy to excimer formation (i.e., WDM) which can, in some cases, be correlated with the temperature coefficient of the solvent v i ~ c o s i t yWhen .~ plotted over a wide temperature range, In ( I f D / I f M ) vs. 1 / T is distinctly curved and actually goes through a maximum; Le., excimer fluorescence intensity is surpressed both at very high and very low temperature^.^^ Thus it should be realized that, in principle, the excimer formation enthalpy ( A H ) and activation barrier ( WDM)can be obtained only in the limits of very high and low temperatures, respectively. Thus the ability of these steadystate data to provide AH and WDMhinges on how far the experimental temperature range is from the maximum in such an Arrhenius plot and also depends upon the relative magnitudes of k M D and k D for a particular excimer system. Discrepancies between enthalpy values obtained from steady-state data (as outlined above) and kinetic data (from AH = WDM - WMD)have been noted b e f ~ r e .An ~ ~Arrhenius *~~ plot of the ratio of excimer to monomer fluorescence intensities is shown in Figure 2 in which the temperature range is wide enough to enable the maximum (at T,,,) to be included. A similar plot for 1-AA over a narrower temperature range is illustrated in Figure 3. It appears to be linear above ca. 40 OC and furnishes an “apparent” AH value of ca. -8.9 kcal/mol which compares with - 10.2 kcal/mol obtained from kinetic measurements. An expression for T,,, can be obtained from eq 1 , assuming that k D , k f D , and k f M are independent of temperature; k M D and k D M are represented as Arrhenius rate constants:
DM = A D MexP(-WDM/RT) Journal of the American Chemical Society
Kinetic Analysis One of the goals in studying excimer formation is to elucidate all of the seven rate constants (and their activation parameters) which pertain to the excimer/monomer scheme outlined above. To achieve this end, one pursues the basic methodology of obtaining and analyzing both the monomer and excimer decay curves as a function both of temperature and concentration. Absolute quantum yield determinations are, of course, necessary in order to partition the monomer and excimer relaxation rate constants (kM and k D ) into their respective radiative and nonradiative components. Solving the appropriate coupled differential equations involving the disappearance rates of excited monomer and excimer results in the following well-known relations for the time dependencies of monomer and excimer fluorescence intensities, respectively: I f M ( t ) 0: exp(-Xlt) + A exp(-Xzt) (2) IfD(f)
= exp(-Xlt) - exp(-XZt)
(3)
where
2 x 1 , ~= x+ Y
d ( x- o2-k 4
k ~ ~ k ( 4~) ~ c
C is the concentration (in M) of the amine, and X , U, and A are defined as: X = kM -/- kDMC (5) Y = kD -/-
k M D -k
kqDC
(6)
Equations 2 and 3 hold for 6 pulse excitation of ground-state amine. Thus it is expected that over certain temperature and concentration ranges, the monomer fluorescence intensity should exhibit a double exponential decay curve. The ratio of the amplitudes of the short to the long components of the monomer decay curve is, then, equal to A. The excimer decay curve is accordingly expected to appear as the difference of two components having equal amplitudes. This decay curve is characterized by the well-known build-up time to maximum intensity which is then followed by an exponential decay (having a slope equal to -XI). Representative decay curves for 1-AA monomer and excimer are shown in Figures 4 and 5 , respectively. Analogous decay curves for the ABCO system have been published e l ~ e w h e r e . ~ Using the following relation for the build-up time, tmax: it is possible to determine X2 from an excimer decay curve (Le., knowing XI) and t,,,. In practice, however, it appears that this
/
99:l
/ January 5, 1977
113 (e.g., a t very low concentrations and/or high temperatures) caused some discrepancy between XI values obtained from an excimer decay curve vis-a-vis the respective monomer decay curve. It is interesting to consider what the limiting values of XI and X2 are as C approaches zero. It is obvious from eq 4 that, with respect to values of XI and X2 calculated for finite concentrations of amine, it is irrelevant as to whether the first term under the square root is expressed as ( X or ( Y - X ) 2 . In fact, it is the latter form which is usually i n d i ~ a t e d .The ~ , ~ limits, however, which are determined for XI and A 2 as C 0 depend on how this term is represented. For example, from eq 4 it follows that
u2
5
t i
I
0
1
i
50
I50
fl
‘
50u \
250
I
1 0
100
300 500 T, N S E C Figure 5. Decay curve of excimer fluorescence of I-AA at 40 centration is 1 .O X M; A,,, is 230 nm.
OC.
Con-
is not a precise way of measuring Xz because of the inherent error in determining tmaxand also the insensitivity of tmaxwith respect to X2. At a fixed concentration and temperature, all directly measurable kinetic information can be represented by XI, X2, and A . In principle, these three parameters can be obtained by a careful analysis ofjust a monomer decay curve, i.e., the two time constants and their relative amplitudes. However, under certain conditions (of concentration and temperature) it is more precise to determine XI directly from an excimer decay curve and to use this value to measure A2 and A by a convolution-and-compare analysis of a monomer decay curve.lx In other cases where both XI, X2 as well as A can be obtained directly from a monomer decay curve, XI derived from the excimer decay curve was used as an internal check of this parameter. Agreement was always found to be satisfactory. However, it appeared that slight spectroscopic “contamination” of the excimer decay curve by monomer fluorescence Halpern et al.
lim X2 = k M
(9)
yet if ( Y - X ) 2 is used, these limits are switched, and XI + k M
-
A2
----*
k~
+~
M D
as C 0. The correct set of limits which applies to an excimer system depends both on the nature of the monomer and excimer and also on the temperature of the system. In other words, if k M > k D k M D limits 8 and 9 are valid; if k M < k D k M D , then limits 10 apply. At ambient temperatures, limits 10 pertain to the pyrene excimer, while limits 8 and 9 describe the saturated amine excimers described in this paper. Because the excimer dissociation rate constant, k M D , is much more strongly temperature dependent than k M and kD,26 there exists a temperature which is characteristic of an excimer system above which limits 10 apply and below which limits 8 and 9 apply. Clearly this “inversion” temperature ( T ’ )depends not only upon the formation enthalpy of the excimer (which, in turn, depends largely on WMD2’), but also on the relative values of k D and k M . This latter condition is determined by the magnitudes of the radiative and nonradiative transition probabilities of the monomer vis-a-vis those of the excimer. It is important to realize that a t T’, the excimer system is nearly in state of dynamic equilibrium. In fact, a t temperatures above this point (in the so-called high-temperature limit), the nonexponentiality of the monomer and excimer decay curves disappears, and both species appear to decay with a common decay ~ o n s t a n t .The ~ , ~utilization of kinetic data obtained in this high-temperature region will be discussed below. In general direct plots of XI and X2 vs. concentration (at a fixed temperature which is > kD and kDMC >> kM to which Ware et add the condition kMD 4- kDMC >> [ 2 ( k ~ ~kDMC)(kD k ~ )’ /]* Equation 16 can also be expressed as
where (K,)o is the molar equilibrium constant pertaining to excimer formation. Accordingly, a n Arrhenius plot of the right-hand-side of (17) can be used to determine A H o and AS”,assuming that both kM and kD are temperature independent. Such an evaluation has been performed for the 1-AA and ABCO excimers using the excimer fluorescence decay data for these amines a t high concentrations M) and a t temperatures between ca. 50 and 100 “C. An Arrhenius plot of the data obtained for the 1-AA excimer is depicted in Figure 16, WherethevaluesofkMand kDusedare 1.31 X lO’and2.2
/ 99:l / January 5, 1977
179 Table V. Thermodynamic Characteristics of the 1-AA and ABCO Excimer Formation. AGO and A H o Are in Units of kcal/mol, and ASo Is in eu Amine excimer ABCO ABCO (vapor) a 1-AA
-AHo
-ASo
(Kdo
-AGO
1.71 x 104 9.76 x 1 0 3 6
5.56
11.6 11.6~
19 20
3.93 x 103
5.0
10.2
17
5.85
stabilization and that therefore the excimer almost certainly does not form in a “head-to-tail” fashion in which the C3 symmetry axes of the molecules are collinear. Excimer formation and the similar thermodynamic properties of 1-AA (vis-a-vis ABCO) also support this idea. A logical configuration to consider for the excimers is the one in which the amines are arranged in a “head-to-head’’ manner (structure I). According to this picture, the two N
Data from ref 10. Value reflects the vapor phase standard state of I atm. Value assumed to be the same as that in n-hexane solution. X 1O6 Hz, respectively. The plot in Figure 16 furnishes values
of A H and A S of -9.0 kcal/mol and -15 eu, respectively, which are consistent with the activation and thermodynamic parameters listed in Tables IV and V. It should be pointed out that AHo and ASoobtained for 1-AA using the method described above are rather sensitive to the particular value of kD used. For example, if kD is taken as 2.0 X lo6, a linear Arrhenius plot is produced which yields a value of AHo of ca. -8.8 kcal/mol. Errors become very large a t low temperatures (ca. 40 OC where X kD) and also a t higher temperatures where X can approach kM. This latter problem did not arise in the case of 1-AA. The reasonable linearity of the In [(kM X)/(X - k ~ ) plots ] implies that for 1-AA both kM and kD are either temperature independent or have only very small dependencies. For ABCO, where kD has been shown to be slightly temperature dependent, this Arrhenius plot was also linear and gave AHoand ASo values which agree well with the data listed in Tables IV and V; namely - 1 1.1 kcal/mol and - 16.4 eu, respectively. It is interesting to consider the possibility of utilizing the Arrhenius plot of eq 17 in elucidating either binding energies and/or kD values for exciplexes and also for weakly bound excimers for which the high-temperature region commences far below ambient temperatures. Presumably for such systems k v is readily obtained. It may also be constructive to use this approach in studying intramolecular excimers and exciplexes in which monomer emission may not be observable except a t very low temperatures.
-
Structure of the Saturated Amine Excimer One of the central questions to be answered concerns the configuration of the saturated amine excimers described in this paper. Although the fluorescence properties of a rather large variety of saturated tertiary amines have been examined, including many acyclic, monocyclic, and bicyclic amines, only the three compounds described here unequivocally undergo intermolecular excimer formation. Thus, even if in other systems where self-quenching occurs and excimer formation might take place, the excimer must be very weakly bound, or otherwise be characterized by an extremely large deactivation rate (relative to the excimers described herein). The criterion for excimer formation and stabilization cannot simply involve the bicyclic nature of these N-bridgehead systems because certain analogous amines, e.g., a derivative of l-azabicyclo [ 3.3.1 Inonane, 1-azabicyclo [ 3.3.31undecane (ABCU), and 1,4-diazabicycl0[2.2.2]octane (DABCO), fail to excimerize. The latter two cases will be discussed below. As mentioned above, the rather large negative entropies of formation for the 1-AA and ABCO excimers (ca. -17 and - 19 eu, respectively) imply a rather tight and discrete configuration. The fact that 4-Me-ABCO excimerizes strongly suggests that the methine hydrogen plays no role in excimer Halpern et al.
I atoms face each other and the molecules are oriented along the same symmetry axis, presumably in such a way so that the bridge systems of the two amines are staggered. This model would apply to all three of the excimer-forming amines. It is also possible that the excimer is not coaxially oriented, but arranged so that one of the N-C bonds of an amine is antiparallel with respect to an N-C bond of its partner as suggested in structure 11. Although there exist other logically possible
0 r---_
I1
structures, the two described above appear to be more intuitively reasonable. On the basis of the thermodynamic and kinetic data alone, it is not possible, of course, to assign or speculate about a particular structure for the excimer. Unlike the case of pyrene, in which the fluorescence properties of the crystal were important in proposing the excimer structure, the respective crystal emission spectra of ABCO and 1-AA provide no definitive information. The crystal emission spectra of both amines are broad and very red-shifted relative to the monomer emission bands observed in solution; the fluorescence, in fact, appears as a violet emission. The crystal structure of ABCO is reported to be face-centered cubic with a unit cell length of ca. 9 A.40,41 The ABCO molecules are presumed to be oriented in parallel arrangements (Le., head-to-tail) in which the “directionality” alternates from row to row. X-ray diffraction studies of the ABCO single crystal appear not to have been reported. It is not clear, then, that the limited information available concerning the ABCO crystal structure can be used to support or even infer the structure of the excimer in the vapor or solution phase. Perhaps the violet emission observed from the ABCO and 1-AA crystals is a consequence of site defects or surface discontinuities.
Nature of the Binding Mechanism The binding forces in excimers and exciplexes are usually considered to arise from two types of interactions: excitation dipole-dipole (exciton resonance) and electrostatic (charge r e ~ o n a n c e ) . ~It? ~is very interesting to consider what the bonding mechanism might be for the saturated amine excimers, and whether the energetics of excimer formation can be rationalized within the framework which is usually applied to the aromatic excimers. In considering purely exciton resonance stabilization, one
/ Kinetics and Thermodynamics of Intermolecular Saturated Amine Excimers
180 must know the electronic transition origins and their respective oscillator strengths. This information is available for ABCO and 1-AA in the vapor phase, although the SI SOtransition for the latter amine has not been measured quantitatively (see Table I and ref 15). The three electronic transitions which are observed for ABCO (below its ionization energy) are weakly to moderately allowed. The oscillator strengths (f values) of the first two transitions for this amine are about 0.003 and 0.06, re~pective1y.l~ The third transition rests upon an underlying continuum making the measurement o f f unreliable. The f value for the S2 SOtransition of 1-AA (shown in Figure 2 of ref 42) is ca. 0.04 which is similar to that of the respective ABCO transition. Simple exciton resonance calculations based on the f values and transition origins of ABCO show that neither electronic transition is strongly enough allowed to provide the required stabilization (i.e., the measured binding energy) a t any reasonable intermolecular separation. Thus, taken alone, exciton resonance appears to be inadequate in rationalizing the excimer binding mechanism. In considering a purely Coulombic interaction between the electronically excited and ground-state species (charge resonance), the stabilization energy (assuming radical cation and radical anion wavefunctions) is expressed as
-
-
WR = I
- A - C(r)
where I and A are, respectively, the ionization potential and electron affinity of the amine, and C(r) is the coulombic attraction of the radical ions (i.e., e 2 / u ) , r being the separation between the (point) charges. Based on the known ionization potentials of the amines (8.02 and 7.94 eV for ABCO and 1-AA, respectively),19 it appears that one might be able to rationalize the stabilization of the amine excimers on the basis of a simple charge resonance model. However, in order to account quantitatively for the observed binding energies of ca. 10 kcal/mol (see Table IV), and assuming a molecular separation of 2.5 8, (e.g., between N atoms in structure I above), the electron affinities of these compounds would have to be >+1.7 eV which seems to be unreasonably large. Thus neither the exciton resonance nor the charge resonance models, considered alone, seems to be adequate in accounting for the stabilization in the saturated amine excimers. It is possible that, like the situation in the aromatic excimers (e.g., pyrene5) configuration interaction between the eigenstates derived from these models might result in excimer wavefunctions whose eigenvalues are more consistent with the experimental data. Irrespective of which model (or models) is (are) used to compose the excimer wavefunctions and eigenvalues, it is crucial that the treatment adequately account for the rather unique specificity which governs saturated amine excimer formation. As mentioned above, no other type of saturated tertiary amine appears to undergo excimer formation, and these include certain analogous compounds such as DABCO and ABCU The nonexcimerization of DABCO can be rationalized on the basis of the fact that this amine is a groundstate coupled system and possesses, therefore, different spectroscopic and photophysical properties relative to monoamines.19b.43-46 ABCU is also rather different from other monoamines because the constraints imposed by the trimethylene bridges cause the partial flattening of the N atom vis-a-vis the three a - C atoms. The C-N-C bond angle is reported to be ca. 11 5 ‘C.)] The possible importance of the C-N-C bond angle, as well as structural rigidity, will be discussed below. Several I N D O calculations were performed on a series of amine pairs in order to elucidate which structural parameters were important in affecting excimer stabilization. The model used for the excited-state amine was the radical cation, a species which represents the limit of electronic excitation for Rydberg states. These calculations indicate that there is a Journal of the American Chemical Society
substantial stabilization for a neutral ABCO molecule in association with the ABCO radical cation (relative to the two neutral molecules) at distances less than about 3 8,. The geometry chosen for this calculation is that indicated in structure I above. Similar calculations were performed for a pair of trimethylamine molecules, one neutral and the other as the radical cation: In case I, the radical was given a planar configuration which is consistent with its known structure as well as that of the excited ~ t a t e , l ~ .and ~ O the neutral molecule was kept pyramidal. In case 11, both species were kept in a pyramidal arrangement, and in case 111, both were constrained to be planar. In each of the three cases, stabilization resulted (relative to the respective pair of neutral molecules), but the magnitude of the stabilization increased in the sequence: case I1 > case I > case 111. It should be realized that in these calculations, the stabilization is to be interpreted qualitatively because the energetics pertaining to the radical cation vis-a-vis the electronically excited states are expected to be different. It is assumed, however, that the relative differences in stabilization energy are significant. Thus, these calculations suggest that it is the ability of the N atom to avoid approaching polarity with respect to the three a-carbon atoms in the excited state which is important in determining excimer stability. This hypothesis can be used to rationalize why ABCU, having a ground-state C-N-C bond angle of about 115 O (which is probably larger in the excited state) fails to excimerize. This rationalization can be likewise applied to the flexible nonexcimerizing amines which are planar, or nearly planar in their excited states. The involvement of some loosely bound excimer in the mechanism of the selfquenching of these amines is, nevertheless, a po~sibility.~’ There is one striking, common structural feature which all three of the excimerizing amines possess. Each of the three a-/3 C-C bonds is directed in parallel orientation with respect to the N n orbital. It is this sort of orbital arrangement which is responsible for the “through-bound’’ ground-state coupling in DABC0.43,44Conceivably, it is this unique configuration of orbitals in ABCO, 1-AA, and 4-Me-ABCO which represents the criterion for excimer formation in the saturated amines. It has been proposed by Hoffmann et al. that the orbital coupling in the transition state of the Grob fragmentation reaction is predicated on a 1-4 interaction in which the N n orbital and the a-/3 C-C (r bonds interact.48 This electronic coupling, then, activates the leaving group in the 4 position, and thus facilitates the fragmentation. If the special structural disposition of the N and a-/3C atoms described above for ABCO and 1-AA were indeed the key factor in promoting excimer stabilization, then one would expect some spectroscopic property of these amines to reflect, accordingly, this situation. Another way of viewing this would be to propose that the aliphatic framework of the cage amine acts as a good electron hole vis-a-vis electronic excitation and that a iarge(er) dipole reversal results consequent to electronic excitation, relative to other amines. If this were actually the case, one would expect that the electronic transition strengths of ABCO and 1-AA would be larger than those of the nonexcimerizing amines. As mentioned above, this is not the case. In fact, there does not seem to be any particularly unusual spectroscopic characteristic of these amines. One (seemingly remote) possibility is that the electron affinities of these cage amines are abnormally larger than those of other saturated tertiary amines. If the I N D O calculations described above are correct in providing some pathway to understanding the structural requirements, a t least, for excimer formation, then one would predict that compounds such as l-azabicyclo[2.2.l]heptane, and other such analogues, whose C-N-C bond angles are constrained by the cage system to be less than ca. 109’ should
/ 99:l / January 5, 1977
181 also undergo excimer formation. In such compounds, of course, the collinear arrangement of the N n orbital and the a-PC-C bonds becomes gradually diminished. Investigations in this area are currently being pursued.
Photochemical Studies Some experiments were performed in order to investigate the consequence of excimer formation on the photochemical stability of the saturated amines. Triethylamine (TEA) and ABCO were both photolyzed in n-hexane solutions, at various concentrations using a 500-W Hg lamp. Degassed solutions of the amines in Suprasil tubes were exposed to the unfiltered lamp for periods proportional to the concentrations of the amines. The disappearance of the starting amine was determined using standard gas chromatographic techniques. Both amines are surprisingly photostable, having disappearance fractions of not more than 0.1 for 50 h of irradiation. For TEA, the disappearance was ca. 3-4% irrespective of concentration (10-4-10-2 M), while that for ABCO appeared to increase from ca. 3% to ca. 10% over the same concentration range and under identical irradiation conditions. These results suggest that the excimer is more photoreactive than the monomer. No other component was observable via VPC consequent to photolysis; however, because of the large amount of starting material present in these analyses, trace amounts of other components could have been masked. In the more concentrated solutions, irradiation resulted in the production of turbid solutions, presumably due to polymer formation. Experimental Section Materials. ABCO was obtained from Aldrich Chemical Co. as the hydrochloride. The free base was liberated by combining very concentrated aqueous solutions of ABCO-HCI and NaOH. The solid base was then filtered and repeatedly sublimed under high vacuum. The first sublimation was usually carried out from a mixture of the base and anhydrous BaO2; ABCO mp, 160 OC (sealed tube). I-AA was provided through the courtesy of Professor W. N. Speckamp. The free base was judged to be adequate for use without further purification on the basis of quantitative absorption spectra, both in the vapor phase and in n-hexane solution (see Table I ) . 4-MA was kindly furnished by Professor C. A. Grob as the hydroperchlorate salt. The free base was liberated using the procedure described above for ABCO. The amine, an oil, wa extracted with a small amount of 2-methylbutane and then retrieved (and dried) via preparative GLC using a Carbowax and KOH-coated Chromosorb P column. All investigations reported in this paper were carried out in n-hexane solution. This solvent (Matheson Spectroquality) was found to be adequate for most measurements without further purification. Certain experiments (e.g., those involving dilute solutions) were more sensitive to background emission, and in these cases, the solvent was purified by being passed through a AgN03eA1203 column and then distilled. Toluene was distilled immediately prior to use, and a,a’-binaphthyl (Eastman) was sublimed just prior to use. Emission Spectra and Quantum Yield Measurements. Fluorescence spectra were obtained using a conventional dc fluorimeter described elsewhere.46 For most experiments, the 232-nm Hg line was isolated from a 200-W Hg-Xe lamp. The excitation band-pass was less than 3nm, and the emission band-pass was usually 2 nm. Emission spectra reported here are uncorrected. Quantum yield measurements of the amine monomer fluorescence utilized toluene (in n-hexane) as the absolute standard. A value of & (toluene) of 0.14 was used. This is the Berlman values0 modified by B i r k ~ . ~For I the amine excimer fluorescence, a,a’-binapbthyl was chosen as the standard because of the favorable coincidence of its fluorescence spectrum vis-a-vis those of the amine excimers. Because of good spectral overlap of the fluorescence standards with their respective amine emission bands,these spectra were not corrected in the quantum yield determinations. Reference and standard solutions were designed such that their respective optical densities at the excitation wavelengths used (238 nm for monomer and 249 nm for excimer) were within about 10% of each
Halpern et al.
other. This was done in order to minimize possible discrepancies arising from optical (geometrical) differences between reference and unknown. In all cases, the observed fluorescence intensities were corrected with respect to the amount of light absorbed. For the measurements of q f M for 1-AA and ABCO, dilute solutions of the amines were used (ca. 1 X M). Although excimer formation is relatively unimportant in these solutions (as evidenced by the apparent absence of excimer emission), excimer formation and dissociation back into monomer was nevertheless taken into account by using the following relation:
where I f M is the (observed) fluorescence intensity of the monomer, and q f M is the intrinsic quantum yield of the monomer. Values of the rate constants were furnished from measurements at ambient temperature. This correction, both for I-AA and ABCO was ca. 10%. For the excimer quantum yield (qfD) measurements, concentrated solutions (ca. 1 X M) of the amines were used. Using the analogous procedure to that described above, the observed fluorescence intensities ( I f D ) were corrected for dissociation into monomer (and, in the case of ABCO, ground-state quenching). These adjustments utilize the following relation:
The estimated errors in the measured quantum yields are 10% for monomer and about 20% for excimer. Fluorescence Lifetime Measurements. The time-correlated single photon method was used; the technique, as well as the particular apparatus employed, has been described elsewhere. For monomer measurements, a 2789 8, interference filter (Corion Instrument Corp.) was used, and for excimer measurements one (or sometimes two) Corning Glass 0-52 filter was used. The flash lamp was a thyratrondriven ( 1 2-20 kHz), low pressure (0.5 atm) D2 discharge constructed from Pt electrodes and which was run at about 7 kV. The fwhm was ca. 2.5 ns, and the l/e time was about 0.8 ns for nearly three decades of decay. For many measurements, a small, high-voltage ceramic disk capacitor (10-20 pF) was added to the firing electrode in order to intensify (and slightly broaden) the main pulse in relation to the long-lived tail. This had the distinct advantage of reducing the distortion otherwise observed in fluorescence decay curves at long times (> ca. 500 ns). In all cases, the monomer and excimer decay curves were analyzed by the convolute-and-compare method using a computer-interfaced X-Y plotter. This rocedure was crucial in assigning precise values to the short-lived decay constant (X2) and relative amplitude of the monomer decay curve. Data acquisition times were kept long enough in order to enable as precise measurements of the decay curve parameters XI, h2, and A as possible. In the most favorablecases, XI and X2 could be measured with less than ca. 1% accuracy. The accuracy attainable in the A values depended strongly upon the particular temperature and amine concentration. Temperature Control. Most kinetic and steady-state measurements were carried out using a jacketed fluorescence cell (Hellma Cells, Inc.) which was attached via a graded seal to a degassing bulb and other associated glassware. Other measurements employed the usual square fluorescence cell which was placed in a machined-out brass block. For high-temperature experiments, ethylene glycol was circulated through the jacketed cell (or the brass block) using a Haake-Ultra-Thermostat. For the low-temperature measurements, a Precision Scientific refrigerated probe was used in conjunction with the temperature bath, in which case methanol was used as the circulating fluid. Temperature accuracy and control was about 0.1-0.5 OC.
Acknowledgment. The authors are grateful to the National Science Foundation for its support of this research. The donors of the Petroleum Research Fund, administered by the American Chemical Society, are also acknowledged for partial support. Professor K. Weiss and Mr. E. Reid were very helpful in the performance of the I N D O calculations. Professors C. A. Grob and W. N. Speckamp are gratefully thanked for their cooperation in providing samples of 4-Me-ABCO and I-AA, respectively.
/ Kinetics and Thermodynamics of Intermolecular Saturated Amine Excimers
182
References and Notes (1) Alfred P. Sloan Research Fellow. (2) Ecole Superieure de Physique et de Chimie lndustrielles de la Ville de Paris. (3) Portions of this material constitute partial fulfillment of the Ph.D. degree of R.J.S.(New York University). (4) J. B. Birks, "Photophysics of Aromatic Molecules", Wiley-lnterscience, London, 1970, pp 301-371. (5) J. B. Birks in "Progress in Reaction Kinetics", Vol. 5, G. Porter, Ed., Pergamon Press, Oxford, 1970. (6) T. Vala, Jr.. I. H. Hillier, S. A. Rice, and J. Jortner. J. Chem. Phys., 44, 23 (1966). (7) H. Lami and G. Laustriat, J. Chem. Phys., 48, 1832 (1968). (8) J. Eisinger, M. Guron, R. G. Shulman, and T.Yamane, hoc.Natl. Acad. Sci. U.S.A., 55, 1015 (1966). (9) A. M. Halpern and E. Maratos, J. Am. Chem. Soc., 84, 8273 (1972). (10) A. M. Halpern, J. Am. Chem. SOC., 96, 4392 (1974). (11) Excimer formation in the noble gases is a well-known phenomenon: see J. B. Birks, Rep. Prog. Phys., 38, 903 (1975). (12) A. M. Halpern and P. P. Chan, J. Am. Chem. SOC.. 97, 2971 (1975). (13) M. B. Robin, "Higher Excited States of Polyatomic Molecules", Vol. 1, Academic Press, New York, N.Y., 1974, pp 208-229. (14) See ref 13, pp 8-31 (15) The electronic absorption spectrum of ABCO has been described by A. M. Halpern, J. L. Roebber, and K. Weiss. J. Chem. Phys., 49, 1348 (1968). (16) A portion of the electronic absorption spectrum of 1-AA has been described by A. M. Halpern, Mol. Photochem., 5, 517 (1973); see ref 19. (17) N. J. Leonard and D. M. Locke, J. Am. Chem. SOC., 77,437 (1955). (18) A. M. Halpern. J. Am. Chem. SOC., 96, 7655 (1974). (19) Consistent with this idea, the vertical ionization potential of 1-AA is slightly lower than that of ABCO: (i.e., 7.94 eV vis-a-vis 8.02 eV): (a) C. Worrell, J. W. Verhoeven, and W. N. Speckamp. Tetrahedron, 30,3525 (1974); (b) P. Bischof, J. A. Hashmall, E. Heilbronner, and V. Hornung, Tetrahedron Lett., 46, 4025 (1969). (20) Reference 13, p 220. (21) A. M. Halpern and D. K. Wong, Chem. Phys. Left., 37, 416 (1976). (22) The radiative rate constants of the monomer and excimer are usually assumed to be independent of temperature: see J. B. Birks, A. J. H. Alwattar, and M. D. Lumb, Chem. Phys. Lett., 11, 89 (1971), and references cited therein. (23) B. Stevens and M. I. Ban, Trans. Faraday Soc., 60, 1515 (1964): B. K.
Selinger, Mol. Photochem., 1, 371 (1969): R. J. McDonald and B. K. Selinger, ibid., 3, 99 (1971). (24) G. E. Johnson, J. Chem. Phys., 61, 3002 (1974). (25) W. R. Ware, D. Watt, and J. D. Holmes, J. Am. Chem. SOC., 96, 7853 (1974). (26) This case seems to apply to most aromatic excimers and certainly pertains to the saturated amine excimers. (27) WMD,the activation energy for excimer dissociation, usually correlates with the binding energy because formation activation energies are usually solvent dependent and smaller than WMo values. (28) This is probably a consequence of error correlations in the XI,? values. (29) A. M. Halpern, D. K. Wong, and A. L. Lyons, Jr., unpublished data. (30) W. C. Danen and R. C. Richard, J. Am. Chem. Soc., 97, 2303 (1975). (31) A. H.J. Wang, R. J. Missavage. S. R. Byrn, and I. C. Paul, J. Am. Chem. Soc., 94, 7100 (1972). (32) Vibronic coupling affected via a high-frequency mode would give rise to a very small temperature dependence. (33) W. R. Ware et al., Abstracts of Vlll International Conference on Photochemistry, Edmonton, Canada, August 1975, and private communication. (34) W. R. Ware and T. Nemzek, J. Chem. Phys., 62, 477 (1975). (35) A. H. Alwattar, M. D. Lumb, and J. B. Birks in "Organic Molecular Photophysics", Vol. I, J. B. Birks, Ed., Wiley, London, 1973. (36) A. Yekta and N. J. Turro, Chem. Phys. Left., 17, 31 (1972). (37) C. Lewis and W. R. Ware, Mol. Photochem., 5, 261 (1973). (38) M.-H. Hui and W. R . Ware, J. Am. Chem. Soc., 98, 4706 (1976). (39) J. B. Birks, A. J. H. Alwattar, and M. D. Lumb, Chem. Phys. Lett., 11, 89 (1971). (40) P. Bruesch, Spectrochim. Acta, 22, 877 (1966). (41) P. Bruesch and Hs. H. Gunthard, fispectrochim. Acta, 22, 877 (1966). (42) A. M. Halpern, Mol. Photochem., 5, 517 (1973). (43) R . Hoffmann, A. Imamura, and J. W. Hehre, J. Am. Chem. Soc., 90, 1499 (1968). (44) E. Heilbronner and K. A. Muszkat, J. Am. Chem. Soc., 92, 3818 (1970). (45) A. M. Halpern, Chem. Phys. Lett., 6, 296 (1970). (46) A. M. Halpern and R. M. Danziger, Chem. Phys. Lett., 16, 72 (1972). (47) A. M. Halpern, D. K. Wong, and B. R. Ramachandran, unpublished results. (48) R. Gleiter, W.-D. Stoner, and R . Hoffmann, Helv. Chim. Acta, 55, 893 (1972). (49) C. Lewis, W. R. Ware, L. J. Doemeny, and T. M. Nemzek, Rev. Sci. Instrum., 44, 107 (1973). (50) I. 8.Berlman, "Handbook of Fluorescence Spectra of Aromatic Molecules", Academic Press, New York. N.Y., 1965. (51) Reference 4, pp 103-105.
A One-Carbon Homologation of Carbonyl Compounds to Carboxylic Acids, Esters, and Amides Stephen E. Dinizo, Robert W. Freerksen, W. Edward Pabst, and David S. Watt* Contribution from the Department of Chemistry, University of Colorado, Boulder, Colorado 80309. Received May 14, 1976
Abstract: An efficient sequence for the one-carbon homologation of aldehydes and ketones to carboxylic acids 2, esters 3, and amides 4 involves ( 1 ) the Horner-Emmons modification of the Wittig reaction using diethyl terr-butoxy(cyano)methylphosphonate (EtO)zPOCH(CN)O-t-Bu to afford a-tert-butoxyacrylonitriles15, (2) the cleavage of the terf-butyl ether in 15 using zinc chloride in refluxing acetic anhydride to afford a-acetoxyacrylonitriles 16, and (3) the hydroxide, alkoxide, or amine solvolysis of 16 to afford 2,3, or 4, respectively, in 57-88% overall yield from the carbonyl compounds.
We have sought to develop useful methodology for the one-carbon homologation of carbonyl compounds 1 to carboxylic acids 2, esters 3, and amides 4. To achieve this versa-
R R"
\c=o 1
-
0
II
Ry
c - z
R'/
\H
2 Z=OH 3, Z=OR" 4,
Z = NR"2
Journal of the American Chemical Society
tility, we required the generation of a transitory acyl intermediate capable of intercepting various nucleophiles to form the desired acyl derivatives. We now wish to report an efficient solution to this problem which relies on the liberation of a masked acyl cyanide.' Available methodology for the transformation of carbonyl compounds 1 to carboxylic acids 2 has relied on the intermediacy of (a) cyanohydrins2 5 (Z = O H ) , (b) nitriles3 5 (Z = H ) , (c) ketene thioacetals4 6 (Z, Z' = SCH3, SCsHs, or S(CH2)3S), (d) a,@-unsaturatedsulfones5 6 (Z = S02C6H5 or S02C6H4-p-CH3; Z' = N H C H O ) , (e) a,@-unsaturated phosphonates6 6 (Z = PO(OC2Hs)z; Z' = N(CH3)2), (f) enol
/ 99:l / January 5,1977
